SAP and Google Cloud are connecting their back-end software so that AI programs can run marketing campaigns without human involvement at each step. A marketing director sets a business goal. The system reads customer data, builds audience groups, writes content variations, sends messages, and adjusts the next send based on who responded. This is what the two companies announced at Google Cloud Next in April 2026.
The tools are not available yet. Marketing is the first use case, and it lands in the second half of 2026. That is a conference announcement, not a shipping date. Enterprise integrations between two large software platforms routinely take longer than planned.
The real problem being addressed is one most businesses already feel. Fewer than two in five companies currently share customer data across their sales or service platforms. A customer calls support, the agent has no record of last week's purchase. A promotional email goes out, the website shows the item as out of stock. These are not edge cases; they are the standard experience. The SAP-Google setup tries to fix this by linking data between the two platforms without physically moving it, so the information stays current and does not have to be duplicated.
For marketing teams, the practical change is significant. Instead of building rigid campaign rules, a team member describes a goal. The AI segments the audience using Google's data analysis tools, generates message content using Google's Gemini AI, and sends those messages through regular channels including SMS and email. When someone responds or ignores a message, the system updates its approach before the next send. No manual reconfiguration required.
There is a larger shift underneath this deal that deserves attention. A shared technical standard called the Universal Commerce Protocol, built by Google together with Shopify, Walmart, Target, Mastercard, Visa, and more than 20 other major companies, is setting the rules for how AI agents interact with retail systems. Under this standard, an AI assistant can check a retailer's inventory, apply a discount, and complete a purchase inside a chat conversation, without the customer ever visiting a website or opening an app.
Consumers are already moving in this direction. A 2026 IBM study found that 45 percent of consumers already use AI for part of their buying process. Around 70 percent say they are at least somewhat comfortable with an AI agent making purchases on their behalf. McKinsey estimates AI agents could facilitate between $3 trillion and $5 trillion in global commerce by 2030.
For businesses that sell products or services, the implication is direct. If your inventory, pricing, and product data are not structured in a way that AI agents can read and act on, those agents will skip your products entirely and send customers to competitors who are set up correctly. BCG puts it plainly: without action, retailers risk being reduced to background utilities in markets controlled by AI agents.
The SAP-Google deal is one response to this. SAP's commerce platform is adopting the Universal Commerce Protocol so that merchant products can surface inside Google Search and Google's AI assistant without requiring retailers to rebuild their existing systems. The AI agent checks live stock before suggesting a product, which means the customer gets an accurate answer and the retailer avoids the classic disaster of a promotional campaign driving demand that the warehouse cannot fulfill.
Google's strategy here is clear. Rather than building a separate marketing suite to compete with Microsoft, Google is making its Gemini AI the engine inside tools that businesses already use: SAP for commerce, Salesforce for customer management, and others. Whether that strategy delivers depends entirely on whether the integrations work as described, and on whether SAP's enterprise customers actually adopt the tools once they are released.
For any business that markets to customers or sells products online, the direction of travel is set. AI will handle more of the routine execution. The question is whether your systems are connected well enough for that to work in your favor.